Spells and Aggro

[TerjynPovar]

I find this subject fascinating, and thus am going to poke my head in.

The first question I have is (and have always wanted to know), is anybody sure that Hate is a pure if X > Y then X has aggro?  I've long wondered if it isn't more like everytime hate changes it's X + Random compared to Y + Random, whichever is more gets it.

The second question I have is did you make a note of Sha's Revenge landing vs. not landing?  I know you said Resists gave just as much hate as landed spells...but perhaps it is modified by a small amount which only gets noticed if the number of resists is high or something similar?

The third and probably weirdest question I have is were you fighting over flat and level terrain?  Proximity in the game is extremely odd, and if you were not the Z-axis issues could have resulted, which could give some extremely odd results which don't seem to fit in with probability.  Heck, with EQ's engine, this is possible even if you are on flat terrain.

Lastly, is it known for certain that casting a spell like the SK's line don't add "detrimental spell cast" hate + the Hate directly from the spell description?  Perhaps this could explain why two 200s was more than a 510.  Although that does seem like a lot, unless you add in an initial value too, which is far more consequential to the 200 hate than the 510 hate.

Unfortunately, none of this is easily testable.  How rare were your statistical outliers anyway?  1 in 3?  1 in 5?  1 in 20?  1 in 100?

[Tastian]

"The first question I have is (and have always wanted to know), is anybody sure that Hate is a pure if X > Y then X has aggro? I've long wondered if it isn't more like everytime hate changes it's X + Random compared to Y + Random, whichever is more gets it. "

Nope, that is one of the many unknowns when it comes to aggro.  In fact like I suggested I think there is a certain bonus for being at the top of the haste list or whatever.  I'm not certain 1 more point of aggro makes you MT instead of someone else.  I honestly hope the RNG doesn't play a part in it and shouldn't imo, but it's possible and trust me I've considered lots of possible explanations.

"The second question I have is did you make a note of Sha's Revenge landing vs. not landing? I know you said Resists gave just as much hate as landed spells...but perhaps it is modified by a small amount which only gets noticed if the number of resists is high or something similar? "

I always noted resist and it was always a case where it was 3 landed or 2 landed + 1 resist.  Or 1 would turn and 1 resist would turn too.  I never had a case where resists increased the number.  Also I had very few resists overall, but I tested some more resisted spells (like the nukes).  This had been tested before too, so it was more of a check to see if it'd changed, none of my data disproved or went againist what was already stated, so I just included it as a way of adding to it.

"The third and probably weirdest question I have is were you fighting over flat and level terrain? Proximity in the game is extremely odd, and if you were not the Z-axis issues could have resulted, which could give some extremely odd results which don't seem to fit in with probability. Heck, with EQ's engine, this is possible even if you are on flat terrain. "

I was over the top careful about this actually.  I made it a point to fight on flat terrain and I kept both chars in third person view so that I could adjust them as I saw fit.  This helped in cases where you could clearly see the SK or bst move back and forth literally a step or two, but aggro would ping pong accordingly.  It also helped me make sure both characters were as relatively close to the mob as possible.  Distance was tested a few times and I have more tests I want to do with distance nuking and distance dot'n, but for now the DoT results really drew my interest and I focused in on those.

"Lastly, is it known for certain that casting a spell like the SK's line don't add "detrimental spell cast" hate + the Hate directly from the spell description? Perhaps this could explain why two 200s was more than a 510. Although that does seem like a lot, unless you add in an initial value too, which is far more consequential to the 200 hate than the 510 hate. "

It's possible, but I have no way of testing because the aggro spells are the only means of hard numbers for aggro we have to go on.  Everythign else in terms of aggro is based upon those very numbers.  If it added hate it's possible that the amount of hate added also scaled based on the amount of hate of the spell so I couldn't even test by juggling 200's and 400's or anything.

"Unfortunately, none of this is easily testable. How rare were your statistical outliers anyway? 1 in 3? 1 in 5? 1 in 20? 1 in 100?"

Outliers for me are far less than 1 in 100.  I literally had over 30 tests of drowsy, 30 of sha's advantage and 40 of revenge at both 200 aggro and 510 aggro.  Then I tested turgurs and got slightly different results than I expected.  I then re-logged bst to re-try my sha's revenge numbers and got a mob that was taking the 2-3 casts to turn 200 hate.  Once that mob died I was back at using 1 cast and 2 for 510.  There was one mob earlier in my testing well doing FoL and listless power that seemed to have the same cling to SK, but not to bst aggro, but I thought maybe I just lost track of a spell somehow or something so I dismissed it.  Overall in hundreds of mobs and multiple tests per mob I'd say I experiened the "freaky" aggro maybe 3 or so times.  Very seldom, but when it happened it was VERY noticeable.  However, it was nothing I could replicate.  I tried pulling same mobs, waiting for mobs to roam to almost the exact same spot, I tried timing, I tried everything I could basically and just couldn't figure it out.  I even messed with distance and had sk further away from bst when the case of "sticky" aggro happened, but it didn't change anything.  All mobs were green, both chars were at 100% life for most of the tests, so it's not like it was frenzy aggro or anything.  

The biggest problem though is we don't have any numbers to go on.  There could be initial aggro cap, might not, might be mod to first aggro, might not.  Basically I'm getting a bunch of numbers and trying to fit them to a formula or explanation, then testing againist my "guess" with the guess for real values I have and seeing if they hold up.  The numbers for SK aggro are the only definite numbers we have though and like you mention even those might be off.  The numbers for melee based aggro come from terror based tests as well.  What we do know is a rough idea of where spells fall relative to each other atleast and relative to terror line's hate #.  Also we have a pretty good idea how healing/nuking and now doting work.  The numbers might not be exact and basically can't be, but the results are very similiar and keep giving the same picture and one that I have been unable to dispprove thus far.  I'm going to take saturday and run a ton of different combinations of dots/focus'd/critical afflictioned at them, but right now it's looking like ~200 "tick" damage is offsetting 200 hate points from SK terror spell.  It's definitely interesting though, so keep firing ideas/thoughts out man.  8)

[Ekss]

Something occurred to me which you might find helpful, since you mention you have a bard: Song of Dawn is a level 53 bard song which lowers hate. Not something I ever used much on my bard, but Lucy does give a range for the amount of hate reduction, and you mentioned that the SK hate spells were the only absolutely numbers we had. There are other hate reduction spells which have 'absolute' numbers too, of course.

It's possible that hate reduction spells don't add any random hate as someone was suggesting the SK spells might - it would hardly make sense for a hate reduction effect also to add hate, but who knows :)

I'm by no means a statistician or someone who's good with numbers or experiments, though, so please excuse me if using hate reduction wouldn't help much.

Have you considered trying to use mem blur to reduce the effect of initial pull aggro in your calculations?

Fascinating topic, by the way! You have me hooked.

[TerjynPovar]

Well, if it was only a couple times Tastian it might just be the Z-axis proximity stuff entering into things, as even when it looks from every angle like you and they are standing right next to each other...

Well, I know for a while that on occasion certain monsters would stand a foot off the ground or so, but the client wouldn't show this, it could only be noticed by doing something that has Z-axis oddities, such as Archery or bolt spells.

Ekss, I'd agree with that it wouldn't make sense for agro reducers to still be some + aggro except for that the Warder procs in effect do this, they have - agro listed in the spell but still are a net + agro, so it's not like it's unheard of.

Good work though...you guys have a patience which amazes me.

[Tastian]

Actually I've used FD with the SK to shed aggro and to wait for aggro wipe "enemies have forgotten you", also done some work with concussion and busted out my ranger for jolt hehe.  Problem is I keep hitting too many unknowns.  For instance can I -hate down to 0 aggro?  I know wizards that open with concussion because they think first spell cast has extra aggro on it so instead of extra aggro on a nuke they toss in a staff click or a concussion then blast away.  

In the case of the outliers it was just VERY noticeable, it was like the sk had a huge aggro modifier or something, I actually kept rechecking spells.  I move them out of range, I had them pull the mob off the bst from a distance, etc.  It's possible it's z-axis or something, but since I can't duplicate it I just don't know.  I had done some initial "distance" aggro tests and had a rough idea of what would pull the mob from outside of melee too, but in this case it was just taking the beastlord 2 to 3 times as much aggro to get things back.  At first I thought I'd just hit a spot in the aggro shifts where it was working out to taking that many, but using higher aggro spells that had shown to have close to if not more aggro than the 200 I couldn't explain it.

One thing I should do is get a 3rd arbitrary character to pull.  I simply don't have a third box anymore though and am down to 2 accounts atm.  You think some people have trouble getting help on groups or epics.  8P  

Also I need to find someway to figure out some values and then build from there.  Like with the dot testing I was able to duplicate the situation very closely which meant that even though I didn't know what the initial aggro was or the value with it, I could take that as ground 0 and from there see the effects of dot ticks/aggro.  I might break out my enchanter and do some mem blur tests, also would let me test their slow for aggro and tash.

Finally, the more I test the more I find quirks that make me think there's atleast a little more to the aggro system than a straight hate list with highest person at the top.  *shrugs*  We'll see what I can come up with this weekend.

[Chackra]

Heals are basically the exact opposite of nukes. It actually does matter how much you heal for. I stood at full life chaining over 12 heals on myself (some crits) and couldn't get the mobs attention. Once I actually healed some damage though I was able to get the mobs attention. This was mentioned awhile ago and the number is somewhere, but I felt was worth throwing in for those that didn't know.

Heal aggro used to be based on potential damage healed instead of actual damage, but they changed this because group heals caused WAY too much aggro that way when most people weren't injured.  Bards used to be famous for getting eaten back when their heal songs caused full aggro even when everyone in the group was at full health.

[Hrann]

Just one thing to throw out there - I know you said your characters were at 100% most of the time, but could the mob be getting a modifier if it realizes it can or cannot damage you?  Like, if it causes good damage to you it has more aggro?

Gerat stuff so far, thanks.

[Coprolith]

Heal aggro used to be based on potential damage healed instead of actual damage, but they changed this because group heals caused WAY too much aggro that way when most people weren't injured.

Didn't they also at some point change it to a fraction of the damage healed? For some reason a 2/3rds modifier just floated to the surface of my memory but i can't remember where i get that number from, its been years since i last played my cleric.

/hugs

[Chubaka]

Just one thing to throw out there - I know you said your characters were at 100% most of the time, but could the mob be getting a modifier if it realizes it can or cannot damage you?  Like, if it causes good damage to you it has more aggro?

Gerat stuff so far, thanks.

Good point.  That would definitly apply to invulnerability, not sure about just because the monster is weak.  We all know that when you get low on hp you will always get your ass beat no matter who is taunting.

I remember WAY back in the day, we (The guild my shaman was in) we used to DA tank (gimpy I know) Vindicator  but then they changed it so that an npc would realize he couldn't hit you, he would go bash someone else.

[Tastian]

"Just one thing to throw out there - I know you said your characters were at 100% most of the time, but could the mob be getting a modifier if it realizes it can or cannot damage you? Like, if it causes good damage to you it has more aggro? "

I thought about this, but couldn't replicate it either.  There are a few aggro mods in my opinion, but with the data I have there's nothing I can pin down atm because everything I find can be explained by half a dozen different things.  

During some of the unslowed tests mobs could do damage to my sk and my beastlord eventually and I had no regen spells or anything going, but it was kinda minor.  However, the damage they were taking was fairly close.  If there were an aggro mod on one char it should have applied about equally to the other.  

There are a few things that just don't look right though either.  The initial aggro thing is weird to me.  Usually if I prox pull something a warrior can drop an aggro and ping the mob off me.  I know I've pulled mobs off of people on incoming with single spells.  Yet in almost all my tests I constantly needed 2 spells to do it.  I think there's just too many unknowns atm to generate reasonable numbers.  Although I do think the aggro on the dots is messed up atm and I don't think there should be any aggro from the ticks because you already pay your aggro when you cast the spell.

[Xarilok]

I know that on my bard, I could do ~55 melee DPS, do group regen, haste, and DS, and have a warrior do ~30DPS, NOT taunt, and pull aggro off me consistantly.

He, however, was lower level, so that may have affected his aggro.

I also know, that I could twist 4 dots, doing over 400 damage per tick, AND debuffing the mob for ~70 Fire, Cold, Poison and Disease, and have a druid do ~270 per tick, with NO debuffs or counters and pull aggro off me consistantly.

The druid was the same level as me.

I have also out melee-dps'd more than one pally or sk, and cast all my usual songs, and w/o them casting or taunting, they have held aggro over my bard time and time again.

I have had other cases where what I was doing SHOULD have locked aggro on my bard hardcore, but didnt.

From what I can tell, some classes have an innate aggro modifier.

[Tastian]

I've actually considered class based aggro mods.   I'm in the process of testing spells that overlap several of my chars.  For instance dots that my nec/bst/shm all have.  I've also been looking over the spell dat and seeing if I can find something or if any of my numbers kinda jump out at me now.  Going to try to finish up the dot testing tonight, then I've got some ranged aggro tests to do, that should help me with pinning down initial aggro.  *shrugs* kinda fun, but kinda annoying too(sounds like EQ doesn't it lol).

[Tastian]

Ok, I've done a bunch more testing and as always I'll try to stay as short and to the point as I can.  The main thing I tested was critical affliction.

Case 1:  Plague - For this I did the usually aggro pull with bst, then have SK pull off with the 200 + 200.  Then I would drop plague with bst.  What happens is it takes initial plague aggro + (2 X 55) to turn the mob the first time to bst, then 3 X 55 to turn the mob the second time.  After that it's 4 X 55 to turn the mob back from 200 hate casts.  This was done dozens of times and always the same deal.  The initial aggro from plague is quite low as we've all known, but just like from previous results it appears that ~200 damage in dot "tick" damage is ~200 aggro.  In the case of plague usually 4 ticks.  However, when critical affliction went off it was clear in the amount of aggro.  Suddenly what would take 4 ticks before to turn the mob took 3.    I kept hoping for back to back crit ticks which I'm certain would turn the mob in 2 ticks instead of 4, but it simply didn't happen.  This on it's own just further went along with most of my previous finding, but after that I went on to...

Case 2:  Scorpion Venom - Scorpion venom usually functions by taking 1 or 2 ticks of the dot (146 damage) to turn the mob from a 200 hate cast.  This was tested well over 200 times.  This also fits with 1 point of "tick" damage being ~1 point of aggro.  It never takes 3 ticks to turn a 200 hate and once a 146 tick happens that turns the mob (whether it's the first or the second) the next cast of 200 hate turns the mob right back.  With critical affliction what happens is that if the mob turns on a 292 "tick" (crit tick) I simply couldn't turn the mob back with a cast of 200 hate.  This too goes along with the 1 to ~1 theory of dot tick aggro.

What I think happens at this point is you cast a dot just like you would any other spell.  When it hits its target it generates aggro.  This happens whether it is resisted or not.  Poison counters make this initial much higher than that of say a disease based dot.  However, once the dot is on the mob the dot continues to generate aggro at about a 1 point of aggro for 1 point of damage ratio.  This appears to be the case regardless of the type of dot and regardless of the cause of the damage.  By regardless of the cause of damage I mean that whether the tick was a "crit" or focused by burning affliction doesn't seem to make a difference.  In my nuke tests the spell itself was what determined aggro.  Having a resisted frost spear or a frost spear that hit for 23 was more aggro than a fully landed crit/focused ice shard.  However, with dots it appears that the actual amount of damage the tick does is the determining factor on the amount of aggro generated.

What I think needs to happen is that aggro needs to be removed from the "tick" portion of dot damage.  You already pay aggro for casting the dot and I see no reason to continually pay more aggro for it every tick it does.

[Caali]

This may be a silly question but when you do all the spell casting tests, do you have your characters face the mob? What I mean by that is you pull aggro and start riposting hits it might account for some of the variations you're seeing. If that's the case you could always just let them beat at your back to be certain you don't riposte.

Good read.

[Oneiromancer]

I am pretty sure Tastian said he faced both characters away and used 3rd person view to see what was going on, for that very reason.

Game on,